Generation of knock-in mice
To deplete male germ cells using HSV-TK, we inserted the P2A self-cleaving peptide and HSV-thymidine kinase (HSV-TK) gene into the C-terminus of the germ cell-specific gene, Vasa (a.k.a. Ddx4 or Mvh) (Fig. 1a). Because male infertility is caused by the short product of HSV-TK that is generated from the testis-specific cryptic promoter in the coding region of HSV-TK,27 we utilized an HSV-TK containing five-point mutations (HSV-TK30).22 Male mice harboring HSV-TK30, unlike regular HSV-TK, have been reported to generate offspring.28
Vasawt/Tk30 and VasaTk30/Tk30 female mice were fertile (Table S1). Male Vasawt/Tk30 mice also produced normal-sized litters (Table S1). In contrast, male VasaTk30/Tk30 mice were infertile, and few, if any, sperm were observed in the epididymis of the VasaTk30/Tk30 mice (Fig. 1b). Furthermore, VasaTk30/Tk30 testes contained a smaller number of elongated spermatids (59.5 elongated spermatids per tubule on average) compared with Vasawt/wt (101.6) and Vasawt/Tk30 testes (103.3) (Fig. 1b, 1c). A close inspection revealed that the heads of the elongated spermatids in the VasaTk30/Tk30 testes seemingly failed to be fully elongate (Fig. 1d). The difference first became clear in stage XII tubules. Consistent with the notion that only the last step of sperm morphogenesis is affected, qPCR analysis did not show any significant differences in the expression levels of different markers between the three genotypes (Fig. 1e).29–36
As homozygous Mvh null mutation has reported to result in arrest of spermatogenesis at the spermatocyte stage,37 the spermatid arrest phenotype of VasaTk30/Tk30 mice suggests that the TK30 locus produces at least partially functional Mvh protein. We considered the following two possible mechanisms for the spermatogenic defects observed in VasaTk30/Tk30 testes: (1) hypomorphic mutation for Mvh functions, (2) deleterious effects of the high level of TK30 expression. To distinguish these two possibilities, we introduced the Mvh-null allele to generate VasaTk30/- mice. No elongated spermatids were observed in VasaTk30/- mouse testes (Fig. 2a). PAS-staining analyses of acrosome formation revealed that round spermatids were arrested at step I or II (Fig. 2b). The qPCR analyses showed that Prm2 expression was decreased in VasaTk30/- mouse testes (Fig. 2c). VasaTk30/- mice showed an intermediate phenotype between VasaTk30/Tk30 and Vasa-/- mice. The results favor the hypomorphic mutation mechanism.
Required dosage of ganciclovir for complete germ cell depletion
To determine the dose of ganciclovir required to successfully deplete germ cells, 3-week-old Vasawt/Tk30 and Vasawt/wt mice were injected with seven different doses of ganciclovir (0.5, 2, 5, 15, 150, 400, 1000 mg/kg, single injection). After 6 weeks, the testes were harvested. In testes injected with 0.5 mg/kg of ganciclovir, no marked difference was observed between Vasawt/wt and Vasawt/Tk30 (Fig. 3a, 3b). In contrast, the injection of 2 mg/kg of ganciclovir in Vasawt/Tk30 mice resulted in significantly smaller and lighter (3.06-fold difference) testes compared with those of Vasawt/wt mice (Fig. 3a, 3b). With the 2 mg/kg dose, spermatids were found in 60.3% of the Vasawt/Tk30 tubules (Fig. 3c, 3d), although spermatocytes were present in mostly all (95.9%) seminiferous tubules. With the 5 mg/kg dose, spermatocytes were found only in 10% of tubules. Remarkably, when Vasawt/Tk30 mice were injected with 15 mg/kg of ganciclovir, more severe effects were observed (Fig. 3a-d). Spermatocytes were present only in 0.55% of the tubules. Almost none of the tubules contained spermatocytes with doses of 150, 400, and 1000 mg/kg (Fig. 3d). However, dramatic decrease in testicular weight was observed even in the control Vasawt/wt testes when high doses of ganciclovir (≥ 400 mg/kg) were injected (Fig. 3b). Spermatocytes were observed only in 14.5% of Vasawt/wt tubules with the 400 mg/kg dose (Fig. 3d, 3e). When 1000 mg/kg was injected into Vasawt/wt mice, all tubules lacked spermatocytes and spermatids. A very high dose of ganciclovir injection is likely harmful to animals.
To determine whether ganciclovir injection similarly affected mature adult mice, 2, 15, and 150 mg/kg ganciclovir was injected into Vasawt/wt and Vasawt/Tk30 mice. The testes were then harvested after 6 weeks. With 2 and 15 mg/kg doses, Vasawt/Tk30 mice exhibited smaller testes compared with the Vasawt/wt mice (Fig. 4a, 4b). However, with the 150 mg/kg dose, not only Vasawt/Tk30 but also Vasawt/wt testes were similarly shrunken. Histological analysis revealed that spermatocytes and spermatids were absent in 90.1% (2 mg/kg), 100% (15 mg/kg), and 98.7 (150 mg/kg) of seminiferous tubules of Vasawt/Tk30 testes (Fig. 4c). When 150 mg/kg of ganciclovir was injected, most (92.2%) tubules did not contain spermatocytes and spermatids even in Vasawt/wt mice.
To examine the long-term recovery of spermatogenesis after the ganciclovir injection, testes were harvested 3 months after injection (15 or 150 mg/kg ganciclovir) into adult testes. In Vasawt/wt mice injected with 150 mg/kg, tubules showing spermatogenesis seemingly more frequently observed in 3 months compared with 6 weeks (Fig. 4b, 4c). Remarkably, almost all tubules of Vasawt/Tk30 mice still lacked spermatocytes and spermatids even with the 15 mg/kg dose. The effect of germ cell depletion is long-lasting in Vasawt/Tk30 mice.
Time course of germ cell loss
To investigate the course of germ cell depletion, adult Vasawt/Tk30 mice were injected with 15 mg/kg of the drug and the testes were harvested at Day 4, Day 14, and Day 25. On Day 4, leptotene spermatocytes were largely depleted in the stage VIII tubules (Fig. 5a, black arrows in the control section show leptotene spermatocytes). Consistent with this, qPCR analyses revealed that Stra8 expression was dramatically reduced (Fig. 5b). In Day 14 Vasawt/Tk30 testes, few or no spermatocytes were observed, whereas all seminiferous tubules contained either round spermatids or elongated spermatids (Fig. 5a). In qPCR analyses, most germ cells genes (Nanos2, Sohlh2, Stra8, Dmc1, PIWIL1) were downregulated, but a spermatid marker Prm2 was unchanged (Fig. 5b). On Day 25, about half of the seminiferous tubules contained elongated spermatids, but other germ cell types were not observed. Decrease in the expression level of Prm2 was observed (Fig. 5b). Sertoli cells were observed in all tubules at all time points (Fig. 5a, b).
Vasawt/Tk30 mice serve as host mice for germ cell transplantation
To examine whether Vasawt/Tk30 mice can be used for germ cell transplantation, cultured SSCs (GS cells) were transplanted into eight testes of Vasawt/Tk30 mice 10–26 days after ganciclovir (15 or 150 mg/kg) injection (Table S2). As the injected GS cells harbor CAG-EGFP transgene (hemizygote),8 successful transplantation can be assessed by EGFP fluorescence. Four months after the injection, four testes were harvested. Three testes showed EGFP signals in large portions (60% or more) (Table S2, Figure S1a). In these three testes, many of the tubules contained spermatocytes/spermatids (Table S2, Figure S1b). These spermatocytes/spermatids were all positive for EGFP fluorescence. In addition, some sperm were observed. The remaining four testes were harvested 9 months after the injection. EGFP was observed in the entire regions (90% or more) in three of the four testes (Fig. 6a, Table S2). In most tubules, normal spermatogenesis was observed (Fig. 6b, Table S2). All germ cells we observed in the 9-month testes were still positive for EGFP.
To determine if sperm derived from the transplanted cells can produce offsprings, testicular sperm were collected from one of the testes harvested 4 months after the injection. Intracytoplasmic sperm injection (ICSI) was performed. On the following day, ~ 41% (9/22) of the injected oocytes developed into 2-cell stage embryos. They were transferred into recipient mice. Two pups were born and one of them showed EGFP fluorescence (Figure S1c). Similarly, epididymal and testicular sperm were collected from the 9-month samples to perform intracytoplasmic sperm injection (ICSI). Approximately 47% (18/38, epididymal sperm) and 42% (8/19, testicular sperm) of oocytes developed into 2-cell stage embryos on the following day of ICSI. These 2-cell stage and 1-cell stage embryos (epididymal, 5; testicular, 1) were transplanted into the recipient oviducts. A total of six pups were obtained (epididymal, 3; testicular, 2; unknown, 1) and four of them were positive for EGFP (epididymal, 2; testicular, 1; unknown, 1) (Fig. 6c). As the transplanted GS cells were hemizygous for CAG-EGFP transgene, most, if not all, were likely derived from the transplanted cells.